I am a researcher on the edge of Natural Language Processing (NLP) and Information Retrieval (IR). My current research focus is on text mining and retrieval in complex domains.
Since 2017, I have been employed as an Assistant Professor at the Leiden Institute of Advanced Computer Science (LIACS). I am affiliated with the Data Science Research Programme of Leiden University. I am group leader of Text Mining and Retrieval Leiden.
I currently supervise projects that develop and evaluate text mining and retrieval methods in a diversity of domains. My group works with a large diversity of textual data: grey literature reports, scientific and legal publications, EU law texts, health records, user-generated content in online patient communities (discussion forums), and news posts on social media.
I collaborate with many different public and private partners such as Sanoma, YoungCapital, the Ministry of Foreign Affairs, The Dutch National Institute for Public Health and the Environment (RIVM), Hogeschool Codarts, and KankerNL.
News and updates
July 2020:
- On July 15, I gave an online training for the Computational Linguistics course in the EU Erasmus+ project “Curriculum Development in Data Science and Artificial Intelligence“
Five bachelor students completed their thesis under my supervision in June and July. (image: “Improving Twitter Stream Filtering using Pseudo-Relevance Feedback” by Bart de Zoete).
- Hugo de Vos has organized an interesting series of industry seminar talks for the TMR group in May, June and July.
June 2020:
- End of semester! I taught a bachelor course (elective) on Data Science with about 30 students. Half-way we had to switch to online teaching and we survived.
- The paper “Helping results assessment by adding explainable elements to the deep relevance matching model” with master student Ioannis Chios was accepted for the workshop on ExplainAble Recommendation and Search (EARS2020) at SIGIR
- Historical linguist Lauren Fonteyn has been awarded a Digital Infrastructure grant for the project MacBERTh, in which she will develop a BERT model for historical English.
- Collega Antal van den Bosch vertelde bij De Taalstaat op Radio 1 over ons project waarin we welbevinden tijdens de coronacrisis meten via sociale-media-analyse.
May 2020:
- Our paper (Alex Brandsen et al.) on creating a NER training dataset in the archaeology domain has been published in the LREC2020 proceeding.
- Our paper (Hugo de Vos and Suzan Verberne) on Challenges of Applying Automatic Speech Recognition for Transcribing EU Parliament Committee Meetings has been published in the Proceedings of the Second ParlaCLARIN Workshop.
- We have published a first draft of a Diversity, Equity and Inclusion (DEI) checklist for ACM-SIGIR events, as a reference for the community
- We have a PhD position open for the project ‘Rethinking news algorithms: nudging users towards diverse news exposure’. The position is at the VU Amsterdam and the first supervisor is Antske Fokkens. I am the second supervisor.
April 2020:
- I am involved in a project on Social Media Monitoring of the RIVM (National Institute for Public Health and the Environment), focussing on the extraction of behavioral aspects related to COVID-19 from open social media. I supervise one of my PhD students (Anne Dirkson) and two master students (Mohamed Barbouch and Rayan Suryadikara) in this project.
- On April 22, we had the annual meeting of the H2020 RISE project ‘Social Media Analytics’, as a virtual online event. I chaired the session about prototypes, addressing the 2020 deliverables for the analysis of social media content.
- All group meetings and thesis presentations have moved online as well.
- Two of my PhD students, Gineke Wiggers and Juan Pablo Bascur, presented their work in the 10th International Workshop on Bibliometric-enhanced Information Retrieval (BIR 2020) on April 14. I recorded a video on invitation by the organizers to congratulate them with the 10th edition of the workshop.
On April 14, I chaired the doctoral consortium at ECIR together with Stefan Rueger. It was a successful online event with two inspiring keynotes.
March 2020:
- Mare published an interview with me about the H2020 RISE project ‘Social Media Analytics’, in which COVID-19 is a relevant current case study.
- An interview about my experience with working and teaching from home at the university’s home page.
- All education has been moved online. My bachelor course has been adapted accordingly with web lectures and interactive sessions in Kaltura: http://tmr.liacs.nl/DS.html
- We had to postpone the MISDOOM 2020 conference because of the COVID-19 situation.
February:
- LIACS organizes BNAIC/BeneLearn 2020 in Leiden. I am one of the local organizers and in charge of the website.
January 2020:
- My group has 7 presentations at CLIN30, 2 talks and 5 posters.
- I am happy to be part of a new interdisciplinary NWO project with Ioulia Ossokina (TU/e), Theo Arentze (TU/e), and Vladimir Karamychev (EUR): BEL (Behaviour, Energy transition, Low income): “Tenants’ behavioural responses to residential energy transition: are intended energy savings feasible?”
- On March 6, 2020 we are organizing the symposium ‘360 Degrees of Data Science‘, giving the floor to Data Science in new and exciting domains.
December 2019:
- I am proud that I am one of the nominees for the faculty of science’s Teacher of the Year award!
- I am currently recruiting two international PhD students in the H2020 project DoSSIER, on Domain-specific Information Extraction and Retrieval. Application web page.
- I am a member of the newly established Ethics Review Committee of the Science Faculty
- I am general co-chair of MISDOOM 2020: the 2nd Multidisciplinary International Symposium on Disinformation in Open Online Media, to be held on April 20-22, 2020 in Leiden
November 2019:
I presented two posters at the Dutch-Belgium Information Retrieval workshop: one together with Benjamin van der Burgh about experiments on Dutch data proving the benefits of ULMFiT for classification with small datasets, and one with Alex Brandsen on the release of the Dutch BERT model, BERT-NL.- My group (in particular Alex Brandsen and Benjamin van der Burgh) has published a number of Dutch-language data sets, together with pre-trained BERT and ULMFiT language models on textdata.nl.
- NWO published the First national research agenda for Artificial Intelligence (AIREA-NL), for which I was a member of the expert committee
- On November 5, I gave an invited lecture in the Lorentz workshop “the future on academic lexicography” on Text Mining for Lexicography. There was a large, interested and engaging crowd. I uploaded the slides here.
October 2019:
- Pre-print just out: “The merits of Universal Language Model Fine-tuning for Small Datasets – a case with Dutch book reviews” by Benjamin van der Burgh and Suzan Verberne on evaluating the effectiveness of ULMFiT for small training sets. Paper on arXiv. Data on GitHub.
- Vacancies! 15 fully-funded PhD Positions on Domain-Specific Search (EU Marie Curie Action project). See the webpage of the DoSSIER project. I am hiring for project 6 and 7, addressing transparency and explainability in legal search. Note that if you have been living in the Netherlands for the last 2 years, you cannot apply for my projects, but you can for those in the other countries.
September 2019:
- On September 10, I was one of the speakers of the event ‘DIT WORDT HET NIEUWS‘, on the future of journalism. My message to the chief editors of three national news papers: Good data journalism is the key
- First Text Mining lecture of the semester with a packed room!
- It is the start of the academic year! We welcome new master students in our group for research projects and thesis projects!
- The course schedule for the Text Mining master course has been published here
August 2019:
- Just published in Multimodal Technologies and Interaction: “Data-Driven Lexical Normalization for Medical Social Media” by Anne Dirkson, Suzan Verberne, Abeed Sarker and Wessel Kraaij
- The result of our interdisciplinary collaboration on junk news was published in PLOS ONE! “The reach of commercially motivated junk news on Facebook” by Peter Burger, Soeradj Kanhai, Alexander Pleijter, Suzan Verberne.
July 2019:
- I presented the paper “Extracting and Matching Patent In-text References to Scientific Publications” (with Ioannis Chios and Jian Wang) in the BIRNDL workshop at SIGIR 2019 in Paris.
- I chaired the Women in IR session during SIGIR 2019 in Paris.
I was interviewed for the Volkskrant, on the topic of Natural Language Processing for Scientific Discovery. “Hoe kunstmatige intelligentie nieuwe kennis opduikt uit miljoenen wetenschappelijke artikelen” (July 19, 2019)
June 2019:
- UMUAI just published our paper Personalized support for well-being at work: an overview of the SWELL project (Wessel Kraaij et al.), the condensed overview of the SWELL project 2011-2017. How to use commodity sensors to interpret behaviour, affect and health status for healthy work-style coaching.
- Our paper “Extracting and matching patent in-text references to scientific publications” (Suzan Verberne, Ioannis Chios and Jian Wang) has been accepted as full paper for the BIRNDL workshop at SIGIR!
On June 18, I was talkshow guest in the Science Café Den Haag, on the theme Big data.- I am honored to be associate editor for ACM Transactions on Information Systems (TOIS)!
- The first Dutch meeting on Clinical NLP (programme) that I organized in Utrecht on June 12, was a big success! Some photos and summaries can be found here.
- On June 11, I was one of the invited speakers in NEMO Kennislink Life, on the theme “Geen nieuws is goed nieuws”
May 2019:
- Our proposal “Constructing a Unified Knowledge Base by joint Deep Learning from images and text” in the Innovation talent programme Leiden-XJTU joint PhD on AI/Bioscience was accepted! This means that Xue Wang will continue her PhD project in LIACS under supervision of Fons Verbeek and me.
- Our paper “Lexical Normalization of User-Generated Medical Forum Data” (Anne Dirkson, Suzan Verberne, and Wessel Kraaij) was accepted for the workshop on Social Media Mining for Health Applications at ACL 2019. In addition, Anne successfully participated in all shared tasks at the SMM4H workshop, resulting in the paper “Transfer learning for health-related Twitter data”.
- Together with a number of LIACS colleagues, I am part of the project Curriculum Development in Data Science and Artificial Intelligence / DS&AI, funded by Erasmus+
- The spring semester is almost finished! I taught my last lecture in the Data Science bachelor course. The final course schedule can be found here.
- I organize an open meeting in Utrecht on clinical NLP on June 12, with 8 presentations (academic and non-academic) and discussion on the challenges of using text data in health records for knowledge extraction and predictive models. Program and registration.
- I am one of the course coordinators of the SIKS course ‘Advances in Information Retrieval’, together with Arjen de Vries and Djoerd Hiemstra. The course takes place on October 8th and 9th in Utrecht.